JP5692117B2 - Rib chip detection device and rib chip detection method - Google Patents

Description

  The present invention relates to a rib chip detection device and a rib chip detection method for detecting a rib chip on a lattice plane of an inspection object.

  2. Description of the Related Art Conventionally, a rib surface is detected by capturing an image of a lattice surface of an inspection object in which a plurality of cells are arranged in a lattice pattern, and performing image processing on the captured image. For example, in Patent Document 1, a monolithic lattice plane in which a regular lattice pattern is formed is imaged, and the presence or absence of a defect in the monolith lattice pattern is determined based on spatial frequency data obtained by Fourier transforming the image. An apparatus is described.

  However, when the inspection object is formed of, for example, a sintered body, the lattice pattern, that is, the shape of the cell may change during sintering. Moreover, the cell located in the outer edge part of a test object differs from the cell in which the shape is located in the center side from the beginning. That is, the inspection object does not necessarily have cells arranged regularly. For this reason, in the inspection based on the premise that the cells are regularly arranged, it is difficult to correctly detect whether or not the ribs are missing. In addition, the inspection object is generally composed of fine cells, and in the case of a sintered body, for example, there may be individual differences in deformation during sintering. It is also difficult to employ a reference sample of interest and to require such a reference sample in advance.

JP 2000-329708 A

  Accordingly, the present invention provides a rib chip detection device capable of detecting a rib chip without requiring a reference sample or the like to be compared even if the test object may have cells that are not regularly arranged. It is another object of the present invention to provide a rib missing detection method.

  The rib defect detection device according to claim 1 is characterized in that the center-of-gravity position of a cell determined to be single by performing a blob analysis on an image captured by an imaging unit that captures an image of a lattice plane of an inspection object Is calculated as an actual center of gravity position, and a reference cell in which no ribs are missing is virtually generated by performing morphology processing on the image, and the center of gravity position of the reference cell is calculated as a reference center of gravity position. Among the reference centroid positions determined that the centroid position and the reference centroid position do not coincide with each other, a midpoint on an imaginary line segment connecting adjacent reference centroid positions is specified as a rib missing position.

  According to this, the center of gravity of a cell determined to be single, more specifically, the center of gravity of a cell determined to be one cell on the image due to the lack of a rib is detected. It is compared with a reference barycentric position that is a reference for performing. The reference center-of-gravity position is calculated based on the image captured by the imaging unit, that is, the image of the inspection object. Therefore, it is not necessary to prepare a reference sample to be compared in advance, and the reference centroid position is calculated for each inspection object, so that it is not affected by individual differences. Further, for example, if the reference center-of-gravity position is the outer edge portion of the inspection object, the center-of-gravity position corresponding to the shape of the cell at that position is calculated as the reference center-of-gravity position. A centroid position corresponding to the shape is calculated as a reference centroid position.

  The reason why the center-of-gravity position does not match when the calculated reference center-of-gravity position is compared with the actual center-of-gravity position is considered to be a state where the actual center-of-gravity position is deviated from the original position due to lack of a rib. Therefore, by specifying the midpoint on the virtual line segment connecting the reference centroid positions as the rib missing position, even if the shape of the cell is different between the outer edge portion and the center side of the inspection object, Even if there is a possibility that the cells are not regularly arranged, it is possible to calculate an appropriate reference barycentric position, so it is possible to detect missing ribs without the need for a reference sample for comparison. can do.

  According to another aspect of the present invention, there is provided a rib missing detection method by performing a real position calculation step of calculating a center of gravity position of a cell determined to be single by performing a blob analysis as an actual center of gravity position, and performing morphology processing. A reference position calculation step of generating a virtual reference cell with no chipping and calculating the center of gravity position of the reference cell as a reference center of gravity position, and determining whether the actual center of gravity position and the reference center of gravity position coincide with each other Extraction of the reference centroid position determined to be inconsistent in the determination step and the determination step, and specify the rib missing position specifying the midpoint on the virtual line segment connecting the adjacent reference centroid positions as the rib missing position And performing a process. Thereby, even if the shape of the cell is different between the outer edge portion and the center side of the inspection object, as in the invention of the rib chipping detection device according to claim 1, the cells are regularly arranged. Even in the case of an inspection object that may not be present, an appropriate reference barycentric position can be calculated, so that a rib defect can be detected without requiring a reference sample to be compared in advance.

The figure which shows schematically the structure of the rib missing detection apparatus by 1st Embodiment. Diagram showing monolithic lattice plane Fig. 2 is an enlarged view of region III The figure which imaged the IV area | region of FIG. The figure which imaged V field of Drawing 2 The figure which imaged the VI area | region of FIG. The figure which shows the flow of the rib missing detection processing by an image processing apparatus Figure 1 schematically showing the procedure for calculating the actual center of gravity position Figure 2 schematically showing the procedure for calculating the actual center of gravity Fig. 3 schematically showing the procedure for calculating the actual center of gravity position Figure 1 schematically showing the procedure for calculating the reference centroid position Fig. 2 schematically showing the procedure for calculating the reference center of gravity position Figure 3 schematically showing the procedure for calculating the reference center of gravity position Figure 4 schematically showing the procedure for calculating the reference center of gravity position The figure which shows the judgment result of the center of gravity position typically The figure which shows the identification result of the rib missing position typically The figure which shows the detection result of the rib chipping position typically The figure which shows typically the detection result of the rib crevice position in a lattice plane An enlarged view of the XIX area of FIG. The figure which shows typically the monolith to test | inspect in 2nd Embodiment. Fig. 1 showing a specific example of rib missing detection Fig. 2 showing a specific example of rib chipping detection Fig. 3 showing a specific example of rib chip detection Fig. 4 showing a specific example of rib missing detection Fig. 5 showing a specific example of rib chip detection

Hereinafter, a rib chip detection device and a rib chip detection method according to a plurality of embodiments of the present invention will be described with reference to the drawings. In a plurality of embodiments described below, substantially the same components are denoted by the same reference numerals, and detailed description thereof is omitted.
(First embodiment)
Hereinafter, a rib chip detection device and a rib chip detection method according to a first embodiment of the present invention will be described with reference to FIGS.

  As shown in FIG. 1, the rib defect detection device 10 according to the first embodiment includes an image processing device 11 and a camera 12, and inspects a monolith 13 that is an inspection object. The image processing apparatus 11 includes a control unit 14, an input unit 15, a storage unit 16, and a display unit 17. The control unit 14 is configured by a computer having a CPU, a RAM, a ROM, an input / output interface, and the like (not shown). The control unit 14 controls the entire rib missing detection device 10 according to a computer program stored in the ROM, the storage unit 16 or the like.

A camera 12 is connected to the input unit 15 of the image processing apparatus 11. The camera 12 has an imaging element such as a so-called CCD or CMOS, and images the monolith 13. In this embodiment, the camera 12 capable of color imaging is employed, but may be capable of monochrome imaging. The camera 12 corresponds to the imaging means described in the claims. The monolith 13 corresponds to the inspection object described in the claims.
The storage unit 16 is configured by a storage medium that can read and write various data, and performs image processing on a computer program for controlling the rib defect detection device 10, an image captured by the camera 12, and the image. Application programs and various data to be executed are also stored. The display unit 17 is, for example, a liquid crystal display, and functions as a display unit that displays an image captured by the camera 12 or a test result described later.

  The control unit 14 includes an actual position calculation unit 18, a reference position calculation unit 19, a determination unit 20, a rib missing position specifying unit 21, and an exclusion unit 22. The actual position calculator 18 corresponds to the actual position calculator described in the claims, the reference position calculator 19 corresponds to the reference position calculator described in the claims, and the determination unit 20 corresponds to the claims. The rib missing position specifying unit 21 corresponds to the rib missing position specifying unit described in the claims, and the excluding unit 22 corresponds to the excluding means described in the claims. In this embodiment, the actual position calculation unit 18, the reference position calculation unit 19, the determination unit 20, the rib missing position specifying unit 21, and the exclusion unit 22 are realized by software by a computer program executed by the CPU of the control unit 14. ing. Note that these units may be realized in hardware.

  Although details of each part will be described later, the actual position calculation unit 18 calculates the center of gravity position of the cell determined to be single as the actual center of gravity position. The reference position calculation unit 19 calculates the position of the center of gravity of the reference cell assuming that there is no missing rib as the position of the center of gravity of the reference cell. The determination unit 20 determines whether or not the actual gravity center position matches the reference gravity center position. The rib missing position specifying unit 21 specifies the rib missing position based on the determination result of the determining unit 20. The exclusion unit 22 will be described in detail in the second embodiment, but excludes a portion of the rib missing position specified by the rib missing position specifying unit 21 that is actually determined to have no rib missing. To do.

  As shown in FIGS. 1 and 2, the monolith 13 is formed in a substantially cylindrical shape and has a plurality of cells penetrating in the axial direction. The end face of the monolith 13 in the axial direction, that is, the end face facing the camera 12 corresponds to the lattice plane. In the case of this embodiment, the monolith 13 is formed of a sintered body such as ceramics. In the monolith 13, as shown in FIG. 3, approximately hexagonal cells 30 are arranged in a lattice pattern, and the cells 30 are partitioned by ribs 31. Here, the direction in which the cells 30 are adjacent to each other, more specifically, the directions indicated by the arrows Y1, Y2, and Y3 (in the case of FIG. 3, three directions) are described in the claims. This corresponds to “cell arrangement direction”.

Next, the operation of the rib chip detection device 10 having the above configuration will be described together with the flow of the rib chip detection method.
In the monolith 13, for example, a part of the rib 31 may be broken during sintering, and thus a rib chip may occur. In FIG. 3, a portion where a rib is missing is shown as a rib missing position P. As shown in FIG. 4, the monolith 13 is different in the shape of the cell 30 (black portion in FIG. 4) between the outer edge portion and the center side. Further, the monolith 13 is deformed as a whole by contraction during sintering, and as shown in FIGS. 5 and 6, the center of gravity X of each cell 30 (indicated by white circles in FIGS. 5 and 6) is outside. A straight line Y (for example, a straight line passing through the center of the cell 30 calculated by the Hough transform) including the cell 30 (see FIG. 5) located on the edge side and the cell 30 (see FIG. 6) located on the center side. May not match. Furthermore, the monolith 13 does not necessarily undergo the same deformation during sintering. For this reason, when the monolith 13 of a sintered body is assumed as an inspection object as in the present embodiment, it is not easy to employ a reference sample applicable to all the monoliths 13.

  Therefore, the rib chip detection device 10 detects the rib chip by executing the rib chip detection process shown in FIG. 7 as described below. In addition, although the rib missing detection processing includes processing executed by each unit such as the actual position calculation unit 18, for example, the control unit 14 will be mainly described for simplification of description.

  The control unit 14 first images the lattice plane of the monolith 13 with the camera 12 (step S1: imaging step). Subsequently, the control unit 14 performs a blob analysis on the captured image, thereby calculating an actual gravity center position R that is a gravity center position of each cell 30 (step S2: actual position calculation step). Here, the blob analysis is a well-known image processing method, and thus a detailed description thereof will be omitted. However, the image is binarized with a predetermined threshold value, and a bright portion and a dark portion (usually a bright portion) of the binarized image. Is a method of analyzing features such as the number, area, and position of a blob (bright part or dark part lump). In the following description, the bright portion of the image is referred to as white and the dark portion is referred to as black. The image data corresponds to data in which white is 1 and black is 0. Further, in FIGS. 8 to 19 described below, reference numerals of the cells 30 and the ribs 31 are omitted.

  FIG. 8 is a binarized image obtained by capturing a part of the lattice plane of the monolith 13. The white area indicates the rib 31 and the black area indicates the cell 30. More specifically, FIG. 8 is an image obtained by imaging a state in which a rib chip is generated in a portion near the center. As a result of binarization, two cells 30 positioned on the left and right of the rib chip position P are connected. Shows the state of one blob. In other portions, no ribs are missing, and one cell 30 is determined as one blob. In this case, the control unit 14 determines that the black set surrounded by the continuous white set is one blob, that is, a single cell 30.

Subsequently, the control unit 14 performs blob analysis on the image shown in FIG. 8 and calculates the center of gravity position of each blob as the actual center of gravity position R. FIG. 9 shows the calculated actual center-of-gravity position R with a white circle symbol. The actual gravity center position R1 shown in FIG. 9 will be described later. When the actual gravity center position R is calculated, the control unit 14 extracts the calculated actual gravity center position R as shown in FIG. In FIGS. 9 and 10, the illustration of the symbol of the actual center-of-gravity position R is partially omitted.
In this way, the control unit 14 calculates the actual center-of-gravity position R by performing blob analysis on the captured image.

  When the actual center-of-gravity position R is calculated, the control unit 14 calculates a reference center-of-gravity position S by performing a morphology process on the captured image (see FIG. 8) in the rib defect detection process illustrated in FIG. S3: Reference position calculation step). Here, the morphology processing is a well-known image processing method, and thus detailed description thereof is omitted. However, for example, a white region included in the binarized image is expanded (increased by 1 pixel) and contracted (1 pixel). Is a process of performing smoothing of an image, removal of isolated points, and the like. In the case of the present embodiment, the control unit 14 expands the white region (that is, the region of the rib 31) in the image of FIG. 8 to virtually show no rib missing as shown in FIG. An image is generated.

Subsequently, the control unit 14 indicates the centroid position in the virtual reference cell in which the black blob in the image illustrated in FIG. 11, that is, the rib lack is not generated, as indicated by the reference centroid position S (indicated by a white circle symbol in FIG. 11). ). The reference gravity center positions S1 and S2 shown in FIG. 11 will be described later. When the reference centroid position S is calculated, the control unit 14 extracts the reference centroid position S as shown in FIG. 12, and generates the masking image shown in FIG. 13 by reversing the extracted image in black and white. In this masking image, the value of the image data is 0 (black) for the reference barycentric position S and 1 (white) for the other values.
As described above, the control unit 14 calculates the center of gravity position as the reference center of gravity position S when it is assumed that no ribs are missing.

When the actual center-of-gravity position R and the reference center-of-gravity position S are calculated, the control unit 14 determines the center-of-gravity position in the rib missing detection process shown in FIG. 7 (step S4: determination step). Specifically, the control unit 14 performs a logical product (AND) operation on the image of the actual centroid position R extraction result (FIG. 10) and the masking image (FIG. 14) generated from the extraction result of the reference centroid position S. . As a result, as shown in FIG. 15, a mismatched centroid position T that is a reference centroid position S that does not match the actual centroid position R is extracted. More specifically, since the actual center-of-gravity position R1 in FIG. 10 deviates from the original position due to the lack of ribs, the portion corresponding to the reference center-of-gravity position S1 shown in FIG. 13 is not masked by the masking image in FIG. The position corresponding to the reference centroid position S1 shown in FIG. 13 is detected as the mismatch centroid position T2 shown in FIG. 15 without being masked by the masking image. Note that FIG. 15 shows a state where the image subjected to the logical product operation is reversed in black and white because of the reference numerals.
In this way, the control unit 14 determines whether or not the actual center-of-gravity position R and the reference center-of-gravity position S match.

  Subsequently, the control unit 14 specifies the rib missing position P in the rib missing detection process shown in FIG. 7 (step S5: rib missing position identifying step). When the reference gravity center position S and the actual gravity center position R are compared, it is considered that the respective gravity center positions do not coincide with each other as described above because the actual gravity center position R is deviated from the original position due to lack of ribs. Further, in such a state, it is considered that a rib defect has occurred between adjacent discrepant barycentric positions T.

Therefore, as shown in FIG. 16, the control unit 14 creates a virtual line segment L <b> 1 between the reference centroid positions S where there is a deviation between adjacent centroid positions T, that is, between the actual centroid positions R. The midpoint on the line segment L1 is specified as the rib missing position P. Then, as shown in FIG. 17, the control unit 14 displays the specified rib missing position P on the actual image of the lattice plane of the monolith 13. In FIG. 17, the actual image is shown in an inverted state because of the reference numerals. Further, the control unit 14 specifies the above-described rib missing position P for the entire lattice surface of the monolith 13, and also displays the rib missing position P on the lattice surface of the monolith 13 as shown in FIG.
Thus, the control unit 14 identifies the rib missing position P by comparing the actual gravity center position R and the reference gravity center position S.

  According to the rib chip detection device 10 and the rib chip detection method described above, the following effects can be obtained.

  The position of the center of gravity of the cell 30 determined to be single by performing blob analysis on the image obtained by capturing the lattice plane of the monolith 13, more specifically, one cell on the image due to the occurrence of a rib defect. The center of gravity position of the cell 30 determined to be 30 is calculated, and a reference cell in which no ribs are missing is virtually generated by performing morphology processing, and the center of gravity position of the reference cell is calculated as the reference center of gravity position S. In this case, it is considered that the actual center of gravity position R does not coincide with the reference center of gravity position S when the actual center of gravity position R is deviated from the original position due to lack of ribs. In addition, it is considered that the rib defect exists between the reference gravity center positions S that do not coincide with each other. Therefore, the actual center of gravity position R and the reference center of gravity position S are compared, and the midpoint on the virtual line segment L1 connecting between the inconsistent center of gravity positions T, that is, between the reference center of gravity positions S determined that the center of gravity positions do not match is determined as the rib missing position. By specifying as P, it is possible to detect a rib chip.

  In this case, since the reference gravity center position S is calculated based on the image of the lattice plane of the monolith 13 to be inspected, the reference sample to be compared in advance is not required, and the sintering is performed at the time of sintering. It is possible to accurately detect the missing ribs of the monolith 13 that may have different contraction modes. Further, since the reference gravity center position S is calculated individually for each examination, it is not affected by individual differences.

  In this case, for example, when a rib chip occurs on the end side of the rib 31, a situation may occur in which the actual rib chip position P is not the midpoint on the line segment L1. However, since the actual rib 31 has a fine structure, it can be said that, from the viewpoint of an inspection apparatus, even if it is determined that the two match each other, it is unlikely that a practical problem will occur. That is, as shown in FIG. 18, it can be said that the function as a device for detecting the rib chipping is sufficiently performed if the position where the rib chipping occurs on the lattice plane is displayed.

  The rib missing position P is specified based on whether or not the actual center of gravity position R calculated based on the image matches the reference center of gravity position S, not the shape of the cell 30 or the rib 31 extracted from the captured image. Therefore, even if the shape of the cell 30 is different between the outer edge of the monolith 13 and the center side, even if the monolith 13 may not be regularly disposed, The missing position P can be specified. Therefore, for example, as shown in FIG. 19, even in a state where ribs are missing between the cells 30 having different shapes on the outer edge of the monolith 13, that is, even in a state where detection is difficult in the prior art, Can be detected.

  As described above, if the rib missing detection method that executes the actual position calculating step, the reference position calculating step, the determining step, and the rib missing position specifying step is employed, the shape of the cell 30 is similar to that of the rib missing detection device 10 described above. Or even if it is the monolith 13 with which the cell 30 may not be regularly arrange | positioned, a rib chip | tip can be detected appropriately.

(Second Embodiment)
Hereinafter, a rib chip detection device and a rib chip detection method according to a second embodiment of the present invention will be described with specific examples with reference to FIGS. 20 to 26. The configuration of the rib chip detection device is the same as that of the first embodiment, and will be described with reference to FIG.
The rib chipping detection device 10 can detect a rib chipping even in a square cell 30 as shown in FIG. The control unit 14 performs the morphing process in the same manner as in the first embodiment, thereby calculating the reference gravity center position S as shown in FIG. Hereinafter, a specific example of the rib lack in the monolith 13 having such a shape will be described.

<Specific example 1>
As shown in FIG. 21A, the first specific example is a state in which the rib 31a between the adjacent cell 30a and the cell 30b is missing. In this case, the control unit 14 calculates the actual center-of-gravity position R as shown in FIG. Then, based on the determination as to whether or not the actual center-of-gravity position R matches the reference center-of-gravity position S, the rib missing position P is specified as shown in FIG.
As described above, the rib chipping detection device 10 can detect the rib chipping similarly to the first embodiment even if the cell 30 is the monolith 13 formed in a square shape.

<Specific example 2>
Specific example 2 includes ribs 31a between adjacent cells 30a and 30b and ribs 31b between adjacent cells 30a and 30c in different arrangement directions, as shown in FIG. This is a state in which chipping has occurred. The control unit 14 calculates the actual center-of-gravity position R as shown in FIG. 22B by performing blob analysis. As a result, three inconsistent barycentric positions T are calculated as shown in FIG. In this case, if the midpoint between the adjacent unmatched centroid positions T is simply specified as the rib missing position P, the intersection of the ribs 31 near the center of the image is also specified as the rib missing position P.

Therefore, the control unit 14 selects between the mismatched centroid positions T adjacent to each other in the arrangement direction of the cells 30 (two directions in the horizontal direction and the vertical direction in the case of FIG. 22) among the mismatched centroid positions T extracted in the vicinity. The midpoint is specified as the rib missing position P. In this case, the arrangement direction can be calculated as a straight line passing through the center of the cell 30 by using the image processing method such as the Hough transform described above. Then, by specifying the midpoint between the mismatched centroid positions T adjacent in the arrangement direction as the rib missing position P, the midpoint between the mismatched centroid positions T not adjacent in the arrangement direction, for example, a rib near the center of the image in FIG. The intersection of 31 is not erroneously specified as the rib missing position P.
As described above, when a plurality of mismatched centroid positions T are detected in the vicinity, the midpoint between the mismatched centroid positions T adjacent to each other in the arrangement direction of the cells 30 is specified as the rib missing position P. Can reduce the risk of detection.

<Specific example 3>
As shown in FIG. 23A, specific example 3 includes ribs 31a between cells 30a and 30b, ribs 31b between cells 30a and 30c, and ribs between cells 30c and 30d. This is a state in which ribs are missing in 31c. In this case, the rib chipping detection device 10 excludes a portion where no rib chipping occurs by the exclusion unit 22. In the case of FIG. 23A, it is determined that the four cells 30a to 30d are single (considered as one blob). As a result, the actual center-of-gravity position R is detected at the position shown in FIG. 23B, and four mismatched center-of-gravity positions T shown in FIG. 23C are extracted as reference center-of-gravity positions S that do not match the actual center-of-gravity position R. The In this case, since four midpoints between the mismatched centroid positions T in each arrangement direction of the cells 30 are detected, four rib missing positions P are specified. However, actually, the rib 31d extending rightward from the intersection of the ribs 31 has no rib chipping.

  Therefore, the exclusion unit 22 is based on which direction the actual center of gravity position R is with respect to the center of the image, in other words, at which position in the region where the actual center of gravity position R is determined to be single by the blob analysis. Based on whether or not there is a portion where a rib chip does not occur is excluded. Specifically, as illustrated in FIG. 23C, the exclusion unit 22 sets virtual lines L <b> 2 and L <b> 3 that pass through the midpoint between the mismatched centroid positions T adjacent in the arrangement direction of the cells 30. The virtual lines L2 and L3 coincide with the positions of the ribs 31 as shown in FIG. In the case of FIG. 23C, the actual gravity center position R is on the left side of the virtual line L2. In this case, it can be seen that the weight on the left side of the virtual line L2 is greater than the weight on the right side of the virtual line L2. That is, it can be determined that the left region has become larger as a result of the lack of ribs on the left side of the virtual line L2.

And the exclusion part 22 excludes the rib missing position P specified on the right side of the virtual line L2 among the four rib missing positions P specified in FIG. As a result, it is possible to exclude a part where the rib chipping is not actually generated, and to specify the rib chipping position P more accurately. Similarly, for example, if the center of gravity position is located to the right of the virtual line L2, the left rib missing position P is excluded, and if the center of gravity position is above the virtual line L3, the lower rib missing position P is determined. If the barycentric position is excluded and below the virtual line L3, the upper rib missing position P is excluded.
As described above, an incorrect portion of the specified rib missing position can be excluded based on where the actual center-of-gravity position is. Therefore, it is possible to detect a rib chip more accurately.

<Specific Example 4>
Specific example 4 assumes monolith 13 shown in FIG. In this case, in the monolith 13, ribs are missing in the rib 31e between the cell 30e and the cell 30f and in the rib 31g between the cell 30g and the cell 30h. In this case, the actual center-of-gravity position is calculated as shown in FIG. 24B by blob analysis. On the other hand, the reference center-of-gravity position S is calculated as shown in FIG. Then, four reference centroid positions S that do not coincide with the actual centroid position R are extracted as shown in FIG. In this state, when the midpoint between the adjacent unmatched centroid positions T is specified as the rib missing position P, the position Pe (center of the rib 31f) indicated by a triangular symbol in FIG. Will be judged. However, in reality, the rib 31f has no rib chipping.

Therefore, the control unit 14 specifies the midpoint as the rib missing position P only between the inconsistent barycenter positions T included in the area U of the cell 30 determined to be single. As a result, as shown in FIG. 24 (E), the rib missing position P is specified in each of the two regions U, while the midpoint between the mismatched centroid positions T deviating from the region U is the rib missing position P. Is not specified as. That is, as shown in FIG. 24F, the rib missing position P can be correctly specified.
As described above, the midpoint between the mismatched centroid positions T included in the region U of the cell 30 determined to be single by the blob analysis, that is, the midpoint between the reference centroid positions S determined that the centroid positions do not match. By specifying the rib missing position P, it is possible to prevent the rib missing from being detected erroneously.

<Specific Example 5>
Specific example 4 assumes monolith 13 shown in FIG. In this case, the monolith 13 has ribs missing in the rib 31i between the cell 30i and the cell 30j and in the rib 31j between the cell 30j and the cell 30k. In this case, the cells 30i to 30k are determined as one blob. For this reason, as shown in FIG. 24, four actual center-of-gravity positions R are calculated, and compared with the reference center-of-gravity position S, two inconsistent center-of-gravity positions T indicated by broken lines are extracted. That is, as a result of coincidence of the actual gravity center position R and the reference gravity center position S of the cell 30j, a state where the mismatch gravity center positions T are not adjacent to each other occurs, and the rib missing position P cannot be specified.

  Therefore, the rib chipping detection device 10 has a reference centroid position S between the mismatched centroid positions T in one blob, that is, the region U of the cell 30 determined to be single by the blob analysis. The reference centroid position S is assumed as a mismatch centroid position T. Then, for each inconsistent barycenter position T in the region U, the midpoint between them is specified as a rib missing position P. Thus, the correct position can be specified as shown in FIG.

(Other embodiments)
The present invention is not limited to the one exemplified in the embodiment, and can be appropriately modified or expanded without departing from the gist thereof.
In each embodiment, the monolith 13 is assumed as the inspection object. However, the inspection object is not limited to the monolith 13 and may be any inspection object having a lattice structure.
The structure of the monolith 13 shown in each embodiment is an example, and other structures such as an octagonal cell 30 may be used. The monolith 13 is not limited to a sintered body, and may be a monolith 13 formed of other materials. That is, the monolith 13 in which the cell 30 is not deformed unlike the sintered body and the shape of the cell 30 is the same throughout the monolith 13 or the cells 30 are regularly arranged may be used.

The actual position calculating step in step S2 and the reference position calculating step in step S3 shown in FIG. 8 may be exchanged in order or may be executed in parallel.
When the masking image shown in FIG. 14 is generated, for example, when the actual centroid position R is extracted (FIG. 10), the extraction result is white (1), or the extraction result obtained by extracting the reference centroid position S (FIG. 13) is used. In the case of white or depending on the arithmetic expression used in the determination process, it is not always necessary to invert the extracted image (FIG. 13).
It is good also as a structure which detects a rib missing part combining the method by each specific example of 2nd Embodiment.

  In the drawings, 10 is a rib missing detection device, 12 is a camera (imaging means), 13 is a monolith (inspection object), 18 is an actual position calculation section (actual position calculation means), and 19 is a reference position calculation section (reference position calculation). Means), 20 is a determination section (determination means), 21 is a rib missing position specifying section (rib missing position specifying means), and 22 is an exclusion section (excluding means).

Claims (6)

Imaging means (12) for capturing an image of the lattice plane of the inspection object (13) in which a plurality of cells (30) are arranged in a lattice pattern;
Real position calculation for performing blob analysis on the lattice plane image captured by the imaging means (12) and calculating the center position of the cell determined to be single in the lattice plane image as the actual center position Means (18);
A reference for performing a morphological process on the image of the lattice plane imaged by the imaging means (12) to generate a virtual reference cell in which no rib is missing, and calculating a centroid position of the reference cell as a reference centroid position Position calculating means (19);
A determination means (20) for determining whether or not the actual gravity center position calculated by the actual position calculation means (18) matches the reference gravity center position calculated by the reference position calculation means (19);
The reference centroid position determined to be inconsistent by the determination means (20) is extracted, and among the extracted reference centroid positions, a midpoint on a virtual line segment connecting adjacent reference centroid positions is defined as a rib. Rib chip position specifying means (21) for specifying the chip position;
A rib chipping detection device comprising:   2. The rib chipping position according to claim 1, wherein the rib chipping position specifying means connects the reference barycentric positions adjacent to each other in the cell arrangement direction with virtual line segments among the extracted reference barycentric positions. Detection device.   2. The rib missing position specifying means specifies a midpoint between the reference centroid positions included in the cell region determined to be single by the blob analysis as the rib missing position. Or the rib chipping detection device according to 2;   Based on the actual center-of-gravity position in the area of the cell determined to be single by the blob analysis, it is determined that no rib chipping has occurred among the rib chipping positions specified by the rib chipping position specifying means. The rib missing detection device according to any one of claims 1 to 3, further comprising an excluding means (22) for excluding the position.   The rib missing position specifying means, if the reference centroid position determined to match the centroid position exists in the area of the cell determined to be single by the blob analysis, 5. The rib chipping detection apparatus according to claim 1, wherein the rib chipping position is specified on the assumption that the reference barycentric position determined to be a reference barycentric position that does not match. 6. Blob analysis is performed on an image obtained by imaging a lattice plane of an inspection object in which a plurality of cells are arranged in a lattice pattern, and the center of gravity of the cell determined to be single in the lattice plane image by the blob analysis An actual position calculating step of calculating the position as an actual center of gravity position;
Morphological processing is performed on an image obtained by imaging a lattice plane of an inspection object in which a plurality of cells are arranged in a lattice pattern to generate a virtual reference cell free from rib defects, and the barycentric position of the reference cell A reference position calculating step for calculating as a reference center of gravity position;
A determination step of determining whether or not the actual gravity center position calculated in the actual position calculation step matches the reference gravity center position calculated in the reference position calculation step;
The reference centroid position determined not to match in the determination step is extracted, and a midpoint on a virtual line segment connecting adjacent reference centroid positions among the extracted reference centroid positions is defined as a rib missing position. Rib chip position specifying step specified as:
The rib missing detection method characterized by performing.

Images